evaluation

Module containing methods for evaluating representations.

This module acts as an interface to evaluation against the classify and retrieval modules through the evaluate_classification() and evaluate_retrieval() functions, respectively.

Author:Kjetil Valle <kjetilva@stud.ntnu.no>
evaluation.evaluate_classification(data, labels, mode='random-split')

Return classification performance using set of data and a list of labels.

Supported values for mode are ‘split’ or ‘cross-validation’.

evaluation.evaluate_retrieval(descriptions, solutions)
Return retrieval performance using cases consisting of descriptions and solutions.

Previous topic

retrieval

Next topic

plotter

This Page